Date of this Version
Evaluation of bulls for calving difficulty of their calves by a multiple trait, mixed model procedure seems appropriate for a number of reasons. Average calving ease scores and their variances and heritabilities differ according to sex of calf and parity of mother. The multiple trait procedure automatically accounts for differences in averages, variances, and heritabilities. The procedure also allows for the magnitude of herd-year-season effects to be different for four traits as defined by calving difficulty scores when measured on: male birth by a heifer, female birth by a heifer, male birth by a cow, and female birth by a cow. Covariances among herd effects for the four traits are used in the multiple trait procedure to prevent certain comparisons within herd from being lost. Records on all four traits are weighted properly to evaluate sires for calving ease for all sex-of-calf and parity-of-cow combinations. Evaluations for the four traits can be weighted to obtain a single evaluation for expected fraction of acceptable births.
Computing costs do not seem prohibitive. The procedure, however, does not take advantage of ordering of categorical responses or of an underlying continuous biological distribution as would threshold models.